Pursuant to the inventive concept discovered by the instant inventors, there is disclosed herein new and improved microcrystalline, wide band gap, n-type semiconductor alloy materials as well as a novel method of fabricating wide band gap n-type and p-type microcrystalline semiconductor alloy materials. It has been synergistically found by said inventors that said semiconductor alloy materials can be fabricated by a glow discharge deposition process utilizing r.f. and/or microwave energy in such a manner as to incorporate a substantial volume percentage of a band gap widening element in the host matrix thereof. Not only is that incorporation accomplished so as to obtain microcrystalline inclusions of a sufficient volume fraction that said materials are characterized by high electrical conductivity, low activation energy and a wide optical gap; but the glow discharge process does not require the presence of high magnetic fields for inducing "electron cyclotron resonance" (as discussed hereinafter).
It is well known in the plasma deposition art that microcrystalline semiconductor alloy material is difficult to fabricate and has heretofore required the use of exotic and complex deposition schemes. In the glow discharge process for the deposition of microcrystalline semiconductor alloy material, it is essential that due to the competing chemical plasma reactions which determine the growth of a film of semiconductor alloy material, vis-a-vis, the etching of that film, great care be taken in the introduction of gaseous precursors. Moreover, if the rate of film growth significantly exceeds the rate of etching, the depositing semiconductor alloy material will not possess a sufficient volume fraction of crystalline inclusions to reach a critical threshold value (as defined hereinafter); and if the rate of etching exceeds the rate of growth, no film will be deposited. It is only when the rate of deposition approximately equals the rate of etching that the required volume fraction of crystalline inclusions will be formed Further, the presence of significant atomic percentages of impurities, such as band gap widening elements, introduced into the plasma has been found to substantially inhibit the formation of crystalline inclusions and has heretofore: (1) prevented the fabrication of n-type microcrystalline semiconductor alloy material characterized by a widened optical gap and high conductivity; and (2) inhibited the fabrication of p-type microcrystalline semiconductor alloy material characterized by a widened optical gap and high conductivity without utilizing a rather complex electron cyclotron resonance technique for the glow discharge deposition thereof
Due to the fact that the instant patent application deals with semiconductor alloy materials which will be referred to by specialized definitions of amorphicity and crystallinity, it is necessary to particularly set forth those specialized definitions at the outset.
The term "amorphous", as used herein, is defined by the instant inventors to include alloys or materials exhibiting long range disorder, although said alloys or materials may exhibit short or intermediate range order or even contain crystalline inclusions. As used herein the term "microcrystalline" is defined by the instant inventors to include a unique class of said amorphous materials, said class characterized by a volume fraction of crystalline inclusions, said volume fraction of inclusions being greater than a threshold value at which the onset of substantial changes in certain key parameters such as electrical conductivity, band gap and absorption constant occurs. It is to be noted that pursuant to the foregoing definitions, the microcrystalline, n-type and p-type semiconductor alloy materials of the instant invention fall within the generic term "amorphous".
The concept of microcrystalline materials exhibiting a threshold volume fraction of crystalline inclusions at which substantial charges in key parameters occur, can be best understood with reference to the percolation model of disordered materials. Percolation theory, as applied to microcrystalline disordered materials, analogizes properties such as the electrical conductivity manifested by those microcrystalline materials, to the percolation of a fluid through a non-homogeneous, semi-permeable medium such as a gravel bed.
Microcrystalline materials are formed of a random network which includes low conductivity, highly disordered regions of material surrounding randomized, highly ordered crystalline inclusions having high electrical conductivity. Once these crystalline inclusions attain a critical volume fraction of the network, (which critical volume will depend, inter alia, upon the size and/or shape and/or orientation of the inclusions), it becomes a statistical probability that said inclusions are sufficiently interconnected so as to provide a low resistance current path through the random network. Therefore, at this critical or threshold volume fraction, the material exhibits a sudden increase in conductivity. This analysis (as described in general terms only relative to electrical conductivity herein) is well known to those skilled in solid state theory and may be similarly applied to describe additional physical properties of microcrystalline semiconductor alloy materials, such as optical gap, absorption constant, etc.
The onset of this critical threshold value for the substantial change in physical properties of microcrystalline materials will depend upon the size, shape and orientation of the particular crystalline inclusions, but is relatively constant for different types of materials. It should be noted that while many materials may be broadly classified as "microcrystalline", those material will not exhibit the properties that the instant inventors have found advantageous of the practice of our invention unless they have a volume fraction of crystalline inclusions which exceeds the threshold value necessary for substantial change. Accordingly, said inventors have defined "microcrystalline materials" to include only those materials which have reached the threshold value. Further note that the shape of the crystalline inclusions is critical to the volume fraction necessary to reach the threshold value. There exist 1-D, 2-D and 3-D models which predict the volume fraction of inclusions necessary to reach the threshold value, these models being dependent on the shape of the crystalline inclusions. For instance, in a 1-D model (which may be analogized to the flow of charge carriers through a thin wire), the volume fraction of inclusions in the amorphous network must be 100% to reach the threshold value. In the 2-D model (which may be viewed as substantially conically shaped inclusions extending through the thickness of the amorphous network), the volume fraction of inclusions in the amorphous network must be about 45% to reach the threshold value. And finally in the 3-D model (which may be viewed as substantially sphericaly shaped inclusions in a sea of amorphous material), the volume fraction of inclusions need only be about 16-19% to reach the threshold value. Therefore, amorphous materials (even materials classified as microcrystalline by others in the field) may include even a high volume percentage of crystalline inclusions without being microcrystalline as that term is defined herein.
Thin film semiconductor alloys have gained growing acceptance as a material from which to fabricate electronic devices such as photovoltaic cells, photoresponsive and photoconductive devices, transistors, diodes, integrated circuits, memory arrays and the like. This is because thin film semiconductor alloys (1) can be manufactured at relatively low cost, (2) possess a wide range of controllable electrical, optical and structural properties and (3) can be deposited to cover relatively large areas. Recently, considerable effort has been expended to develop systems and processes for depositing thin film semiconductor alloy materials which encompass relatively large areas and which can be doped so as to form p-type and n-type semiconductor alloy material. Among the investigated semiconductor alloy materials of the greatest significance are the silicon, germanium and silicon-germanium based alloys. Such semiconductor alloys have been the subject of a continuing development effort on the part of the assignee of the instant invention, said alloys being utilized and investigated as possible candidates from which to fabricate thin films of amorphous, microcrystalline and also polycrystalline material.
Multiple stacked cells have been employed to enhance photovoltaic device efficiency. Essentially, and pursuant to the concept of stacking cells in electrical optical series relationship, the cells are fabricated from different band gap semiconductor alloy material, each cell adapted to more efficiently collect various portions of the solar spectrum and to increase open circuit voltage (Voc). The tandem cell device (by definition) has two or more cells with the light directed serially through each cell. In the first cell a large band gap material absorbs only the short wavelength light, while in subsequent cells smaller band gap materials absorb the longer wavelengths of light which pass through the first cell. By substantially matching the generated currents from each cell, the overall open circuit voltage is the sum of the open circuit voltage of each cell, while the short circuit current thereof remains substantially constant.
It is now possible to manufacture high quality layers of intrinsic thin film semiconductor alloy material utilizing techniques developed by the assignee of the instant invention. However, the n-type layers of thin film semiconductor alloy material heretofore fabricated have, in many instances, been of less than the optimum quality required for the manufacture of the highest efficiency electronic devices therefrom. Accordingly, and because of the limitations imposed, inter-alia, by the n-type semiconductor alloy material, the optimum operational potential of many classes of thin film semiconductor alloy devices, such as solar cells, have as yet to be achieved. The instant inventors have further found that even the p-type, wide band gap, microcrystalline semiconductor alloy material which was the subject of the patent grant in U.S. Pat. No. 4,609,771, failed to optimize the performance of photovoltaic devices incorporating said material.
For example, if a highly transparent, wide band gap, highly conductive, microcrystalline, n-type semiconductor alloy layer (also referred to as a highly "n-type layer") and a highly transparent, wide band gap, highly conductive, microcrystalline, p-type semiconductor alloy layer also referred to as a highly "p-type layer") could be fabricated, p-i-n and n-i-p type photovoltaic cells manufactured with said wide band gap microcrystalline, n-type and p-type semiconductor alloy layers would exhibit not only significant, but synergistic improvement in the operational performance thereof. Such highly n-type and p-type layers would be characterized by low activation energy and would thus increase the magnitude of the electrical field established across the layer of intrinsic semiconductor alloy material, thereby improving the fill factor of the photovoltaic cell fabricated therefrom. Similarly, the built-in potential of the photovoltaic cell, and consequently, the open circuit voltage generated thereacross would be increased by the presence of the highly conductive, very wide band gap layers of p-type and n-type microcrystalline semiconductor alloy material. Also, since the built-in potential of the cell is increased charge carriers are more readily moved from the photoactive region in which they are photogenerated to the respective electrodes despite the presence of photoinduced defects which are responsible for the well known effect of Staebler/Wronski degradation, thereby providing drastically improved stability. The improved electrical conductivity exhibited by the microcrystalline n-type semiconductor alloy material, vis-a-vis similarly constituted and doped amorphous semiconductor alloy materials, which materials are characterized by a number of crystalline inclusions below the aforementioned threshold value, would further provide for decreased series resistance encountered by charge carriers in their movement through the photovoltaic cell. The decrease in series resistance would result in an improved fill factor and an improved overall efficiency of that photovoltaic cell.
Furthermore, wide band gap layers of n-type microcrystalline semiconductor alloy material and even wider band gap layers of p-type microcrystalline semiconductor alloy material are more optically transparent than corresponding layers of amorphous semiconductor alloy material or corresponding layers of doped microcrystalline semiconductor alloy material which do not incorporate band gap widening elements. Such transparency is desirable, if not essential, not only on the light incident doped layer, but also in the junction layers of a stacked p-i-n type photovoltaic cell because increased transparency allows more light, whether incident upon the p-type layer or the n-type layer or redirected by a back reflector through those p-type and n-type layers, to pass therethrough for absorption in the layer of intrinsic semiconductor alloy material (the photogenerative region) of the lower photovoltaic cell. It is in the photoactive layers of intrinsic semiconductor alloy material that charge carrier pairs are most efficiently photogenerated and separated. Therefore, photovoltaic cells employing microcrystalline, wide band gap, n-type layers and microcrystalline wide band gap p-type layers of semiconductor alloy material would also produce higher short circuit currents due to the more efficient collection of shorter wavelength, highly energetic and blue and green light in the intrinsic material thereof and the resultant photogeneration of charge carriers therefrom. This consideration of transparency would, of course, increase in significance as the number of stacked, individual p-i-n type photovoltaic cells in a tandem device increases. This is because, in such a tandem photovoltaic device, a light absorbing (narrow band gap) n-type layer in (1) the upper photovoltaic cell would "shade" one or more of the underlying cells and thus reduce the amount of incident light absorbed in the intrinsic semiconductor alloy layer, the layer with the longest lifetime for charge carriers photogenerated therein, and (2) the lower photovoltaic cell would "shade" one or more of the superposed cells and thus reduce the amount of light redirected by a back reflector absorbed in the intrinsic semiconductor aloy layers. Obviously, with respect to the light incident p-type layer, the more the band gap can be widened without decreasing the conductivity thereof below about 1 ohm.sup.-1 cm.sup.-1, the less incident radiation is absorbed therein and the greater the open circuit voltage which can be generated.
No reported technique has disclosed a method of fabricating wide band gap microcrystalline n-type silicon alloy material to which have been added a substantial atomic percentage of an impurity, such as a band gap widening element. The only technique for the incorporation of a band gap widening element into a host matrix so as to fabricate wide band gap microcrystalline p-type silicon alloy materials is disclosed by Hattori, et al in a paper entitled, "High Conductive Wide Band Gap P-Type a-SiCH Prepared By ECR CVD And Its Application To High Efficiency a-Si Basis Solar Cells" presented at the 19th IEEE Photovoltaic Conference on May 4-8, 1987. As described therein, highly conductive p-type, wide optical gap, microcrystalline silicon alloy material was prepared utilizing microwave power in a chamber about which a magnetic flux is established of about 875 Gauss. The ECR (electron cyclotron resonance) plasma is extracted from the ECR excitation chamber and moved into the deposition chamber along with a gradient of dispersed magnetic field. The extracted ECR plasma interacts with the reaction gas mixture of SiH.sub.4, CH.sub.4 and B.sub.2 H.sub.6 so as to promote the growth of a p-type a-SiCH film characterized by an optical gap of, for instance 2.25 eV and a dark conductivity of about 10 ohm.sup.1- cm.sup.-1. There was no disclosure presented by Hattori, et al as to the volume fraction of crystallites or the size of those crystallites; however, it may be assumed (based upon the conductivity and band gap) that the volume fraction necessary for the aforementioned percolation theory criteria has been satisfied.
This work compares favorably with the results reported herein, wherein the resultant p-type, wide band gap silicon alloy material is characterized by an optical gap of at least 2.1 eV, a dark conductivity of about at least 1 ohm.sup.-1 cm.sup.-1, an activation energy of about 0.05 eV and microcrystalline inclusions amounting to at least 70 volume % in the amorphous network. Both the work of the instant inventors and the work of Hattori, at al may be utilized to fabricate the p-type layers of semiconductor alloy material in solar cells, especially the aforementioned tandem solar cell structures, as to provide synergistic increases in photoconversion efficiencies; which synergistic increases were described in commonly assigned U.S. Pat. Nos. 4,600,801 and 4,609,711. However, a very important difference between the work of the instant inventors and the work of Hattori, et al must be noted.
The difference resides in the fact that the subject inventors are able to fabricate the wide band gap, p-type (as well as n-type) microcrystalline semiconductor alloy material by a process which does not rely upon an exotic process such as the electron cyclotron resonance process of Hattori, et al in order to obtain microcrystalline inclusions in a semiconductor alloy material which contains a significant atomic percentage of a band gap widening element, such as carbon, nitrogen or oxygen. More specifically, it is well known that the fabrication of microcrystalline silicon alloy material becomes exceedingly difficult as the atomic percentage of modifiers such as carbon, nitrogen or oxygen are added thereto. (See for example, the article in the JournaI of Non-Crystalline Solids 59 & 60 (1983) pp. 791-794, Hiraki, et al entitled Transformation of Microcrystalline State of Hydrogenated Silicon to Amorphous One Due to Presence of More Electronegative Impurities, which article found that as little as 2 molecular percent of N.sub.2 provides for a microcrystalline silicon hydrogen film to change into an amorphous state.) It is for precisely that reason Hattori, et al was compelled to employ electron cyclotron resonance in a microwave plasma system in order to grow crystallites in an alloy which included only 10 volume percent of carbon. As a matter of fact, it must be noted that the conductivity of the film of Hattori, et al was drastically reduced as the high microwave frequency was reduced to the r.f. range.(see FIG. 8 of the Hattori, et al paper). This provides clear evidence that Hattori, et al required the presence of both a high magnetic field and a highly energetic microwave plasma in order to deposit a silicon alloy material characterized by the desired value of conductivity, optical gap and activation energy.
In mark contrast thereto, the subject inventors have significantly simplified the fabrication processes of Hattori, et al and have been able to achieve the aforementioned characteristics of high conductivity, wide optical gap and low activation energy without resorting to elaborate and expensive production techniques, such as electron cyclotron resonance. Moreover, the subject inventors have developed an r.f. glow discharge process for the deposition of such wide band gap, p-type, microcrystalline silicon alloy material, which process does not require the presence of a magnetic field at all; much less a magnetic field sufficient to induce electron cyclotron resonance. In order to stress the importance of this development, it is informative to note that there is no disclosure (other than that of Hattori, et al) of which the instant inventors are aware, in either the scientific or patent literature, of the production of wide band gap, microcrystalline silicon alloy material to which an impurity (such as a band gap widening element) has been added, whether that material is n-type or p-type. This is because the narrow regime in which to deposit microcrystalline semiconductor alloy material referred to hereinabove, which regime exists between the deposition of amorphous silicon alloy material and the etching of that material, is further reduced when significant atomic percentages of impurities, such as nitrogen, carbon or oxygen are introduced into the plasma.
Therefore, the instant inventors are the first to have developed a simple process by which to deposit n-type and p-type microcrystalline semiconductor alloy material to which band gap widening elements have been added As a matter of fact the instant inventors are the first to have developed an n-type microcrystalline wide band gap semiconductor alloy material. These and other objects and advantages of the instant invention will become apparent from a perusal of the Detailed Description Of The Invention, the Brief Description Of The Drawings and the Claims which follow.